Positron emission tomography (PET) was used for the study of regional cerebral perfusion and metabolism in eight patients with severe post-hypoxic encephalopathy, caused by cardiac arrest and resulting in a coma lasting for at least 24 h. Using this method, we aimed to identify regional vulnerability, which was hypothesized to provide (i) insight in pathogenic mechanisms and (ii) early prognostic parameters. On day 1 post-resuscitation, 18-Fluor deoxyglucose ([F18]-FDG) indicated a marked decrease of cerebral metabolic activity. Gray matter glucose consumption was 54% of normal values, whereas white matter uptake was 70% of normal. Regional differences followed a pattern of neuronal density rather than specific patterns of functionally or biochemically defined regions or of vascular territories. In contrast to [F18]-FDG, the distribution of 15-oxygen labeled water ([O-15]-water) showed a better demarcation between gray and white matter, whereas focal deficit was not observed. In some patients, hyperperfusion relative to regional glucose consumption was observed in the occipital poles and basal ganglia. This suggests loss of vascular tone, i.e. vascular paralysis, in the basilar artery territory. CT and MRI scanning did not show any major change with respect to the hypoxic injury. In the small group studied, all patients had a poor outcome. The comparison between survivors and nonsurvivors did not reveal obvious differences in PET data, suggesting that this technique does not provide major prognostic clues adding to the prognostic information derived from serial neurological assessment in the restricted patient group characterized by prolonged coma.